The purpose of this proposal is to develop faster, safer methods of endovascular embolization of cranial tumors by intraoperative tracking and display of the catheter within a three-dimensional (3D), tree-based map of the patient's vasculature. Endovascular surgery of the cranial vasculature is difficult. Intraoperatively, the interventionalist threads his catheter through a complex vascular tree, guided only by x-ray angiography (angiography). Puffs of contrast medium are injected through the catheter, highlighting on projection view a local region of a vascular tree. No 3D information is available. Instead, the clinician uses only 2D projections, monitors catheter advancement almost continuously, and frequently changes the angle of view in the attempt to visualize 3D anatomy. The procedure is therefore lengthy, requires multiple angiograms, and commonly results in an hour or more of radiation exposure to the patient. This proposal aims to reduce the difficulty of the procedure and the immense radiation exposure delivered to the patient by allowing the interventionalist to interpret each 2D, intraoperative angiogram within a 3D context. There are 3 specific aims: a) To create a 3D, tree-based map of each patient's vascular anatomy. This map may be created from segmented MRA data alone or may include the additional detail provided by reconstructing preoperative angiographic data into 3D. b) To create a preoperative planning environment, in which the clinician may view a tree-based description of the vasculature from any interactively selected angle of view and in relationship to the target tumor. One or more optimal catheter pathways may be selected and saved for later intraoperative use. The program will run on standard hardware and under a common operating system, allowing study of the vasculature on readily available computers or at home. c) To register the 3D vascular map with each intraoperative angiogram, allowing the interventionalist to visualize, in 3D, t he position of the catheter within the proper subtree. Both the location of the tumor and a preselected optimal catheter pathway will also be shown in 3D. Each 2D angiogram can thus be interpreted within its 3D context, diminishing the difficulty of the procedure, the number of angiograms required, and the patient's radiation exposure and contrast load. We include a prospective, randomized, controlled clinical trial of the efficacy of our methods in both reducing radiation exposure to the patient and in providing more effective embolization. The methods to be developed are immediately applicable to tumor embolization and chemotherapy delivery anywhere in the body. They are also applicable to many thousands of other endovascular procedures preformed each year.